Follow up future plan-review note: Consider supporting custom semantic similarity thresholds or matchers for fact dedupl

[wwind123]

Future follow-up from approved planning for issue #535.

Source context:

• Parent issue: Reconcile and curate evidence in final discuss summaries #535
• Approved plan subject: e9692d871b1c32d476df3d3f31f9e3541139131a7658aea9d7a557c7ec13082b
• Approved plan hash: e9692d871b1c32d4
• Planning round(s): 1, 2
• Reviewers: Antigravity, Claude
• Original plan item ID(s): item-7, item-9

Canonical follow-up:

• Consider supporting custom semantic similarity thresholds or matchers for fact deduplication in the future.

Original reviewer notes:

• item-7, round 1, Antigravity: Consider supporting custom semantic similarity thresholds or matchers for fact deduplication in the future.
  • Update from Claude: Still independent; the current plan uses an LLM-based optional reconciler for paraphrase grouping rather than configurable similarity thresholds, so this remains valid future work.
  • Update from Antigravity: Consider supporting custom semantic similarity thresholds or matchers for fact deduplication in the future.
• item-9, round 2, Claude: Consider supporting custom semantic similarity thresholds or matchers for fact deduplication in the future.

This was approved as future work during planning. It is outside the current implementation scope and is not a PR-review prior item.